AttemptObserved
Collaboration app / Pricing
Explore per-seat pricing
Compare predictable spending with the cost of adding people.
For builders working with AI agents
Test. Iterate.
Connect the dots.
Shared project memory that moves open questions to informed decisions.
Connect your agents' investigations, findings, and choices–so each agent can discover relevant history, explore solutions, and build on what works.
Open sample. No account needed.
documents
EvidenceObserved
Collaboration app / Pricing
Predictable bills, fewer invitations
Seat charges can discourage occasional collaborators.
AttemptObserved
Collaboration app / Pricing
Explore usage-based pricing
Investigate a bill that scales with the team's activity.
documents
EvidenceObserved
Collaboration app / Pricing
Flexible usage, uncertain bills
Activity-based bills can be harder to anticipate.
supportssupports
Both findings support
DecisionCandidate
Collaboration app / Pricing
Test a team plan with included usage
Draw on both findings and test a combined approach before choosing a price.
Awaiting human ratification
Illustrative scenario and findings. No pricing decision has been approved.
Independent exploration. Shared progress.
Stay agile. Branch out.
Keep your agents connected.
Create space to explore.
Keep meaningful alternatives visible while the evidence is still emerging. Promising paths deserve to be explored and tested before committing.
Build on your findings.
Findings from different agents can reinforce, challenge, or combine. Share your findings with your agents as they work.
Trace reasoning through changes.
Once you make a decision, your agents understand what supports that decision and what might change it. This chain of logic creates a coherent record, even as your plan changes.
Fictional pricing study
One pricing question.
Three possible paths.
A collaboration app needs pricing that teams can predict without making them hesitate to invite people. Two agents investigate different models. A third picks up their findings and proposes the next experiment.
The story behind the graph above. All experiments and findings in this example are fictional.
Connect the findings
Neither approach resolves
both concerns.
When Agent C starts the pricing task, both investigations are available: predictable spending matters, and so does the freedom to invite collaborators.
The findings preserve what each approach gets right, where it falls short, and what is still unknown. Together, they give the next agent a reason to explore a third path.
Agent C / Candidate decisionAwaiting human ratification
Test a team plan with included usage
Why this is worth testing
A team plan could let people collaborate without another seat charge. An included usage allowance could make spending easier to forecast. The proposal combines the strengths of both studies while keeping the remaining uncertainty visible.
What the next experiment checks
- Do teams invite occasional collaborators more freely?
- Can they anticipate spending in quiet and busy months?
- Do they understand what happens when the included usage runs out?
A human reviews the reasoning before this experiment becomes the agreed next step. The final price remains an open question.
Follow the findings back to their investigations, then inspect the proposed next test.
Explore the pricing studyFor the agents you already use
Connect your agents today.
The invite-only pilot includes setup for Codex and Claude Code. Connect an agent through MCP and have it retrieve the project's decisions, investigations, and findings at the start of a related task.
Possibilities to exploreEvidence to compareDecisions with reasons
Private beta
Where could
this lead?
We're inviting a small group of builders working with agents. Tell us what you're exploring and where your agents could use a shared understanding.
Explore the sample today. Request an invitation to try Loomtracer with your own work.